Search and the web
A simulated search run on your behalf by a tool. It is not what your customer sees, and the distance between the two is the whole of what a ranking report leaves out.
Also called Position tracking · Отслеживание позиций · Keyword ranking
A tracker queries the engine from its own infrastructure, in a chosen country and language, on a chosen device, with no history and no account. Your customer queries it from their own city, signed in, on their phone, having searched three related things this week. Personalisation, location down to the neighbourhood, device, and the searcher's own history all move the result, and none of them exists in the tracker.
Two consequences follow and both are routinely omitted from reports. A tracked position is a sample from a distribution rather than a fact — the same keyword measured twice in one day from two datacentres can differ by several places. And the position counts the ten blue links only: a listing at rank three sitting under an AI overview, a map pack and four ads is at the top of a list nobody reaches.
None of this makes tracking useless. It is the only way to watch a specific query over time under fixed conditions, and fixed conditions are what make a change interpretable. It stops being useful the moment the fixed conditions are dropped from the report.
It is a fact about a machine in a datacentre with no history, on a device and in a location somebody chose when the project was set up. Ask which location, which device, and whether the figure is an average across several checks. Then ask what sits above the result on the page — because the honest version of a top-three ranking under an AI overview is that the query is answered before your listing is reached.
Knowing the definition is not the same as being able to check the figure. These are the procedures that do the second thing.
- Testing a ranking report
- “You now rank in the top ten for 47 keywords, up from 12.” · 25 minutes, 5 questions.
The definitions are the easy part. Whether the figure on your dashboard was computed this way is a different question, and usually the more expensive one.